MIT Researchers Built an App that Detects Emotion in Conversation

emotion-detecting app

The researchers hope that the emotion-detecting app will help individuals with anxiety disorders, autism, or Asperger’s in the future.

Depending on the way an individual tells a story, those who are part of the conversation can draw multiple conclusions. For example, when talking about a recently purchased car can come across as either brag or excitement, depending on the listener. To get rid of such uncertainty, a team of researchers at the Massachusetts Institute of Technology designed a smartphone app able to pick up on certain indicators of emotion behind each part of a story.

Emotion-Detecting App

For their research, scientists at MIT used a fitness tracker. The app, which is built into the device, collects speech and physical data in real time while also analyzing the overall tone a person uses to tell the story. So far, with the help of artificial intelligence, the application can tell the difference between sad and happy parts of the story, tracking emotional changes in five-second intervals.


Multiple participants to the study were asked to wear the fitness tracker provided by Samsung, more specifically, the Samsung Simband with the built-in emotion-detecting app and asked to tell a story. As the subjects were speaking, the app began to record any physical changes such as heart rate, fidgeting, skin temperature, and other movements like waving of the arms.

The study, although unclear if peer-reviewed, deemed the application 83 percent accurate when determining the overall tone used by the story-teller. According to the researchers, the artificial intelligence linked parts of the speech that used monotonous tones or had long pauses to sadness. On the other hand, varied speech patterns were associated with happiness.


Tuka Alhanai, graduate student at MIT, believes that even though the concept is still in its infancy, at one point people will be able to rewind conversations and see the exact moments when individuals around them felt most anxious during a conversation.

In the future, it is possible that individuals could have their own AI social coach in their pockets, adds Tuka Alhanai. She believes that the app will prove most helpful to those with anxiety disorders or conditions such as autism or Asperger’s. Also, this is not the first time MIT researchers took an interest in analyzing human emotions. The latest innovation is part of an ongoing effort the Computer Science and Artificial Intelligence Laboratory branch in studying human interaction through emotion.

Image Source: Flickr

About Waleed Javed